> > From: "Dennis Brown" <see3d@xxx <see3d%40writeme.com>>
> > To: <
amibroker@xxxxxxxxxxxxxxx <amibroker%40yahoogroups.com>>
> > Sent: Saturday, January 05, 2008 5:55 PM
> > Subject: Re: [amibroker] Hello Tomasz, Which C++ compiler do you
use to
> > compile AmiBroker ?
> >
> > > Tomasz,
> > >
> > > Wow, a man after my own heart!
> > >
> > > It is rare today to find a programmer that understands the
machine
> > > from the ground up. We have so many levels of abstraction
today. I
> > > have speed tested some of your array operations and have been
very
> > > impressed with them. Now I know how you did it --you spent a
> > > lifetime preparing.
> > >
> > > I started a bit earlier (1970) --having to design my own TTL
CPUs
> > > before LSI microprocessors. I progressed along the same lines,
> > > though I had easy access to parts, but an actual computer was
beyond
> > > my means. I was motivated by a desire for a "personal" computer
and
> > > a love of programming (productivity tools) --which I only had
access
> > > to at the local colleges. I had to design and build my own:
> > > 8 bit serial TTL computer -- coded with port-a-punch cards and 8
> > > LEDs for output
> > > 8 bit 8008 microprocessor -- coded through a teletype paper
tape reader
> > > 12 bit parallel TTL computer -- programmed through a teletype
paper
> > > tape reader
> > > 16 bit virtual memory TTL computer -- moved up to high speed
paper
> > > tape reader CRT and line printer
> > > 6800 microprocessor "Jupiter II" computer (I made kits for
sale).
> > > Finally Floppy Disks
> > > Z80 microprocessor "Jupiter III" computer. Finally a hard drive.
> > >
> > > All of these were programmed at the machine code level and my
own
> > > programming was generally for implementation of simple
interpreter
> > > languages (in my day "BASIC" would be considered high level)
and RT
> > > direct control of automated machine tool motors. Software DDA
> > > algorithms for stepper motors and machine operator UI and
control
> > > languages etc.
> > >
> > > By 1980 I was specializing in floppy disk drive designs then
hard
> > > disk drives and I never designed another CPU. Though the hard
disk
> > > drives had a 68k CPU system on a chip and were more powerful
than any
> > > computer I had designed in a big box!
> > >
> > > In 1984 I purchased my first computer designed by someone else -
- an
> > > Apple Macintosh. I went from needing to build them to just
wanting
> > > to use them to do other things.
> > >
> > > I had a disk drive array project that I wanted to use a
Transputer
> > > product in. It was quite an interesting product, though the
project
> > > did not get funded.
> > >
> > > In the late 90's and up to a couple of years ago I was writing
my own
> > > EOD stock market simulation and graphing applications in
HyperCard/
> > > SuperCard/Revolution. However, when I wanted to jump to intraday
> > > trading, I went looking for a faster program. Higher level
languages
> > > like Revolution running on a VM are great for creating custom
user
> > > interfaces and general problem solutions, but RT markets
require real
> > > purpose-built array processing. AmiBroker was the fastest
program I
> > > found and a great price too.
> > >
> > > If Apple had not moved to the intel processor, I might never
have
> > > looked at AmiBroker (I am spoiled by OS X). The world is indeed
> > > shrinking and differences are quickly disappearing.
> > >
> > > Just as I tipped my hat to Apple for making a better PC than I
did, I
> > > tip my hat to you for making a better market analysis program
than I
> > > did. It frees me to use rather than invent solutions.
> > >
> > > Thank you,
> > > Dennis
> > >
> > >
> > > On Jan 5, 2008, at 5:47 AM, Tomasz Janeczko wrote:
> > >
> > >> Hello,
> > >>
> > >>>> Microsoft Visual C++ 6.0 SP6 for 32 bit version and Visual
C++
> > >>>> 2005 for 64
> > >>>> bit version.
> > >>>
> > >>> TJ, are you aware of
http://agner.org/optimize/ - might be
useful
> > >>> for you ...
> > >>
> > >> Thanks, but with all respect due, I don't think so.
> > >> As a 10 year old I started learning about TTL logic (gates,
> > >> counters, de/multiplexers, shifters)
> > >> moved later to arthmetic logic, 12 year old I built among other
> > >> things 16-step digital sequencer for
> > >> analog synthesiser.
> > >> as a 13 year old I started my coding experience with hand
coding in
> > >> machine 8080 code (directly in binary, without help of
> > >> assembler),
> > >> then coded entire OS for my own microcomputer based on Zilog
Z80
> > >> (when I was 17)
> > >> then coded Motorola MC68000 demos in m68k assembler (20 years
old)
> > >> and critical parts of AmiBroker code on PC are also optimized
> > >> on assembly level. My masters thesis involved assembly on INMOS
> > >> T8xx Transputers
> > >> (anyone here knows what it was?)
> > >> In the past wrote native code emulators, cross compilers and
> > >> many low-level programs all involving hand optimizations, hand
> > >> register allocations,
> > >> manual CPU cycle calculation, etc.
> > >> With all that background I can safely say that I can write
> > >> optimized code already.
> > >>
> > >> Best regards,
> > >> Tomasz Janeczko
> > >>
amibroker.com
> > >> ----- Original Message -----